It can call tools or APIs
Start with an AI Agent Security Assessment or MCP Security Review when the system can change real workflows, data, tickets, code, or deployments.
- AI Agent Security Assessment
- MCP Security Review
Review selector
Match the review to what the system can do and the decision you need to make.
The right review depends on authority: what the system can read, write, delete, send, execute, retrieve, approve, or remember.
Overview
Compare reviews by system access and purpose.
- A guide to choosing a review
- Recommendations based on system permissions and the decision you need to make
- Links to the most relevant assessment, certification, briefing, or retainer page
- Help defining the work when more than one review applies
01
Choose by system permissions
It retrieves sensitive content
Start with a RAG Security Assessment when isolation, authorization, chunk handling, source labels, or generated responses can expose protected materials.
- RAG Security Assessment
It writes code or influences delivery
Use an AI coding or CI/CD review when the system can create pull requests, run commands, touch secrets, alter workflows, or influence deployment authority.
- AI Coding Security Review
- AI-assisted CI/CD Risk Matrix
Leadership needs context first
Use a briefing or retainer to prepare for a board discussion, evaluate vendor claims, or get recurring technical advice.
- Executive and Board Briefing
- Advisory Retainer
02
Choose by review purpose
- Production launch: AI Agent Security Assessment or AgentBoundary readiness.
- Enterprise procurement: Enterprise Buyer Packet plus review package.
- Certification preparation: AgentBoundary Certification Readiness.
- Board or investor decision: Executive and Board Briefing.
- Ongoing roadmap support: Advisory Retainer.
Next step